Aqueous synthesis of high surface area metal oxides

Catalysis Today 137 (2008) 132–143

Contents lists available at ScienceDirect

Catalysis Today journal homepage: www.elsevier.com/locate/cattod

Aqueous synthesis of high surface area metal oxides Christian Bluthardt b, Carola Fink a, Klemens Flick b, Alfred Hagemeyer a,*, Marco Schlichter a, Anthony Volpe Jr.a a b

Symyx Technologies Inc., 3100 Central Expressway, Santa Clara, CA 95051, USA University of Applied Science, Max-Planck-Street 39, D-74081 Heilbronn, Germany

A R T I C L E I N F O

A B S T R A C T

Keywords: High surface area Metal oxides High throughput Combinatorial Pechini Combusion synthesis Catalysts Absobents Carriers Supports

By applying high throughput synthesis and characterization technologies, we have been optimizing common dry or aqueous synthetic routes for the preparation of high surface area metals and oxides, such as precipitation and modified Pechini methods. For wet combustion synthesis, we have been screening a variety of organic acids as dispersants and developed proprietary recipes for individual metals. By resorting to easily decomposable organic acids (as opposed to citric acid in the original Pechini combustion method), such as glyoxylic acid, oxalacetic acid and ketoglutaric acid, it is possible to obtain high surface area materials for many metals after careful optimization of acid/metal ratio and calcination conditions. Examples are Sn, In, Co, Ru, Ni, Fe, Mn, Y, Ce and Rare Earth oxides and their mixtures. After calcination in the temperature range of about 300–400 8C, surface areas >150 m2/g could be obtained for Er, Tm, Co, Ru, and Nb; >200 m2/g for Sn, Fe, Mn, and Y; >300 m2/g for Ce; and >400 m2/g for Ni oxide. Noteworthy are also >140 m2/g for La2O3, >80 m2/g for CuO, and 75 m2/g for ZnO. For V, around 40 m2/g was possible for the nearly carbon-free V2O5, whereas up to 90 m2/g was obtained for a 90% V–10% carbon composite (by incomplete combustion of the organic acid). Residual carbon helps in stabilizing the porous oxide against sintering. Thus, conventional aqueous routes (precipitation, Pechini) can be competitive to more elaborate and costly methods such as those using organic solvents, sol–gel, supercritical drying or template/hydrothermal synthesis. Combustion synthesis is well suited for the preparation of mixed oxides from mixed metal solutions in aqueous organic acids. Bulk porous Co and CoRu mixed oxides have been screened for liquid phase alcohol oxidations and CoRuCe oxides for CO oxidation and VOC destruction, and doped NiO has been reduced to the metal and tested for various hydrogenations. ß 2008 Elsevier B.V. All rights reserved.

1. Introduction High surface area metal oxides are desirable absorbents, carriers and catalysts, and various synthetic methods starting from aqueous metal solutions or slurries have been applied. These include precipitation, hydrothermal synthesis, microemulsions as well as wet combustion synthesis. Auxiliary reagents such as acids or bases, templates, surfactants or dispersing agents are added to convert, immobilize or finely divide the metal precursors that upon subsequent drying and calcination are transformed to the corresponding oxides. Organic solvent assisted routes such as solvothermal synthesis, sol–gel with or without supercritical drying, or oil drop methods are also widely used for the preparation of high surface area materials. The various traditional synthetic methods that are available for the

* Corresponding author. E-mail address: [email protected] (A. Hagemeyer). 0920-5861/$ – see front matter ß 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.cattod.2008.04.045

synthesis of inorganic nanoparticles have been extensively reviewed and evaluated [1,2]. Recently many more sophisticated methods for the preparation of nanoparticulated materials have been proposed, e.g. flame pyrolysis, arc and plasma discharge, sputtering, microwave irradiation, molten salt flux methods, solid state reactions, mechanical alloying, and inert gas condensation. We have investigated the potential of three different traditional (dry or aqueous) routes, namely dry decomposition, precipitation and combustion synthesis from aqueous solutions, for the synthesis of high surface area metals and oxides starting from common and readily available metal precursors without making use of expensive templates, surfactants or alcoholic solvents or supercritical drying or high pressure equipment. For liquid-phase combustion synthesis, we have been screening a variety of organic acids as dispersants and developed proprietary recipes for individual metals. Combustion synthesis is well suited for the preparation of mixed oxides from mixed metal solutions in aqueous organic acids. By resorting to easily decomposable organic acids (as opposed to citric acid in the original Pechini combustion

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

method [1–5] or malic acid in a later modification [6]), such as glyoxylic acid and ketoglutaric acid, it was possible to obtain high surface area materials for many metals after careful optimization of acid/metal ratio and calcination conditions [7–18], for example, Sn, In oxides and their mixtures [8,9,12,14], yttria [15] and ceria [18]. Some catalytic applications of these high surface area metal oxides have already been published, for example, SnO2 as carrier for emissions control [12], ceria as carrier for water–gas-shift catalysts [19], and Co–Ru mixed oxides for CO and total oxidations [20,21]. We have now applied and optimized combustion synthesis methods for additional metals such as Ni, Co, Fe, and Rare Earths and compared surface areas to those obtained by precipitation. We will summarize previous and recent results, classify metals according to most suitable acid dispersant and corresponding calcinations conditions and point out common trends.

133

centrifugation in two 40 ml vials followed by 4 wash steps in 40 ml vials. A typical temperature ramp was as follows: 55 8C/4 h ramp/ 120 8C/4 h hold/1 h ramp/350 8C/4 h hold in 40 ml vial. 2.1.2. Combustion synthesis Libraries consisting of arrays of stirred tall 40 ml glass vials were loaded by powder or liquid dispensing robots with solid metal precursors and/or metal solutions as well as aqueous solutions of organic dispersants and/or solid organics, optionally aged at room temperature in open vials to evaporate part of the solvent (to induce gelation but also to prevent splashing and excessive foaming during the subsequent drying step), then dried and calcined in air in static calcination ovens. Amounts and process conditions used are detailed in the following tables. 2.2. Synthesis recipes

2. Experimental 2.1. Equipment 2.1.1. Precipitation Our integrated co-precipitation workflow has been published elsewhere [17,18]. Briefly, the workflow consists of an automated co-precipitation station, an automated particle recovery station, ovens for thermal treatment, and a parallel sample agglomeration and sizing system for particle sizing. The Symyx Software suite is used to design the, control the robotic actions and store the resulting data, and view the data. The co-precipitation station consists of a robot with eight parallel precipitation channels and 8 columns of 8 precipitation wells resulting in an 8  6 array of wells. Each channel has a pH probe, temperature probe, and three liquid addition probes that are robotically inserted into a disposable liner with magnetically coupled stirring. The Station is shown in Fig. 1. The precipitation can be carried out as a constant liquid addition, as a titration, or as a constant pH precipitation. Precipitation or pH titration recipes were then scaled up in 250 ml three-neck flasks equipped with masterflex pH controllers and peristaltic pumps. Isolation of the precipitate was done by

2.2.1. Dry decomposition Common metal salts such as oxalate, acetate, carbonate, hydroxide were purchased from Aldrich, Alfa or Pfaltz&Bauer and used as supplied. Metal formates were prepared by slurrying metal carbonate or hydroxide in excess formic acid followed by solvent evaporation. The dry powders were calcined in air in tall 40 ml vials according to the following representative heat up protocol: 55 8C/1 h ramp/120 8C/1 h hold/1 h ramp/325 8C/4 h hold. 2.2.2. Wet combustion synthesis A general recipe consists of dissolving or slurrying the metal precursor (acetate, nitrate, hydroxide) in aqueous organic acid (glyoxylic acid OHC-COOH, ketoglutaric acid or acetone-1,3dicarboxylic acid HOOC-CH2-CO-CH2-COOH, oxalacetic acid HOOC-CH2-CO-COOH) in a tall 40 ml glass vial followed by calcination in air. A typical heat up protocol is as follows: 55 8C/ 4 h ramp/120 8C/4 h hold/1 h ramp/300 8C/4 h hold. 2.2.3. Synthesis of Ni–Ce–Pd by the modified Pechini method 1 g Ni hydroxide was dissolved in 6.5 ml 20% aq. glyoxylic acid in a 40 ml glass vial, then appropriate amounts of solid Ce acetate and ammonium Pd oxalate were added under stirring to obtain a

Fig. 1. Co-precipitation station and supporting catalyst synthesis hardware.

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

134

clear green solution that was calcined at 320 8C for 4 h in air in a static calcinations oven using the standard heat up ramp mentioned above in 2.2. Ce contents varied from 1 to 6 wt% while Pd contents were in the range 0.02–0.2 wt%. 2.3. Characterization BET surface area, N2 adsorption isotherms and pores size distributions were measured in a three-channel Micromeritics Tristar 3000 porosity analyzer. Chemical composition and carbon content was analyzed by Galbraith Labs. XRD was recorded on a PANalytical X’Pert Pro and SEM on a Hitachi S4300. 2.4. Integrated workflow and sample throughput We have focussed on three realistic and therefore commercially relevant synthetic methods and have chosen the screening tools accordingly (co-precipitation robots for the precipitated samples, liquid dispense robots for the metal-acid solutions, and powder dispense robots for the solids). It should be stressed that even the oxides prepared by dry decomposition required a high throughput approach, namely solids handling by powder dispense robots. Hundreds of samples need to be handled and tracked which would be impossible without an integrated workflow. The libraries are, of course, simple arrays of vials. No optimization algorithm or evolutional approach was taken and the multidimensional

parameter space mapped out by a straight forward grid search with orthogonal libraries (two gradients per library). The calcination is very efficient as many arrays of vials can be processed in parallel at the same time in large static ovens. The bottleneck of the workflow was the three-channel BET instrument, where approximately 20 samples could be characterized per day. The libraries (arrays of vials) were manually transferred to the calcination ovens. This step was not automated but was still not the bottleneck since many arrays could be processed in parallel (e.g. a stack of 5 plates of 48 samples each). The workflow was accelerated in many cases by eliminating bad samples based on their color (e.g. holmium should turn pink) and weight (mass balance indicative of complete combustion of organic dispersant). In summary, the time needed to characterize a 24-sample library is roughly 1 day. The level of reproducibility is approximately 10%. BET strongly depends on the calcination temperature (which therefore must be optimized in small increments of 5–10 K steps), and reproducibility can be poor when using calcination ovens with slightly different actual temperatures from the set value and/or with temperature distributions. 2.5. Catalytic tests NiO samples were pre-reduced at 250 8C by a continuous flow of 10% H2 in N2 and then tested for catalytic activity in the

Table 1 High surface area NiO–Ni precursor screening Precursor

State

Calcination

BET (m2/g)

Yield (mg)

Theory (NiO) (mg)

500 mg Ni(OH)2

Grn dry pwd

300 8C/4 h 275 8C/4 h 250 8C/4 h

140 171 177

393 390 398

403 403 403

500 mg Ni carbonate 500 mg Ni acetate 500 mg Ni hydroxyacetate

Grn dry pwd Grn dry pwd

300 8C/4 h 300 8C/4 h 300 8C/4 h

141 104 173

310 152 180

315 179

Recalcined Recalcined

300 8C/1 h 300 8C/1 h 300 8C/0.5 h

72 240 267

382 249 194

199 199 199

Recalcined Recalcined

275 8C/4 h 275 8C/2 h 275 8C/2 h

197 227 260

237 191 177

179 179 179

Grn dry pwd Grn dry pwd Grn dry pwd Bluish dry pwd

300 8C/4 h 300 8C/4 h 300 8C/4 h 325 8C/2 h

10 216 135 410

169 138 120 204

202

500 mg Ni oxalate

Bluish dry pwd Recalcined Recalcined

325 8C/1 h 325 8C/0.5 h 325 8C/0.5 h

414

200 mg 400 mg 500 mg 500 mg

Red Red Red Red

300 8C/4 h 280 8C/4 h 260 8C/4 h 320 8C/4 h

112 33 6 66

43 99 145 106

310 8C/6 h 325 8C/2 h

273 382

389 296

555 mg Ni hydroxyacetate

500 mg Ni hydroxyacetate

500 mg 500 mg 500 mg 500 mg

Ni Ni Ni Ni

Ni Ni Ni Ni

formate citrate lactate oxalate

dimethylglyoxime dimethylglyoxime dimethylglyoxime dimethylglyoxime

700 mg Ni oxalate

pwd pwd pwd pwd

Bluish pwd Recalcined

121 204

No decomposition, still bluish Partial decomposition, partly black 216 204

286

Ni hydroxide = Ni(OH)2 Alfa 12517 FW92.72 Ni carbonate = NiCO3 Alfa 22897 FW118.72 Ni acetate = Ni(OOCCH3)2xH2O Alfa 12223 Ni hydroxyacetate Ni(OOCCH2OH)2 Alfa 39456 FW208.75 Ni formate = Ni(OOCH)22H2O Alfa 13801 FW184.77 Ni citrate = Ni3(C6H5O7)2xH2O Alfa 39634 Ni lactate = Ni(OOCCHOHCH3)24H2O Alfa B23643 FW308.91 Ni oxalate = NiC2O42H2O Alfa 39454 FW 182.76 Ni dimethylglyoxime = Ni(C4H7N2O2)2 Alfa 39458 FW288.94 Thermal decomposition of the dry powder (as supplied by Alfa). Standard temperature ramp: 45 8C/4 h ramp/120 8C/4 h hold/2 h ramp/300 8C/4 h hold.

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

hydrogenation of octene under the reaction conditions T = 55 8C, m(cat) = 300 mg, V(C8H16) = 1.5 ml, V(CH3OH) = 50 ml. 3. Results Some representative results for Ni, Fe, Rare Earth as well as Cobased mixed oxides will be presented below. 3.1. Results for nickel – Ni precursor screening – dry decomposition Combustion synthesis has already been applied to Ni oxide but only low surface areas <100 m2/g have been reported ([22] and references therein). Commercially available NiO nanopowder (Alfa 44297, 0.008–0.01 mm APS powder) had a BET surface area of 83.6 m2/g (as supplied, no pretreatment, no outgassing) or 87.2 m2/g (outgassed at 250 8C overnight). Commercial micronsized NiO (Aldrich 24,403-1; <10 mm particles) had a BET surface area of 31 m2/g (as supplied). We have investigated the thermal decomposition of dry powders of Ni salts that are convertible into the oxide upon air calcinations. As seen from Table 1, high surface areas >400 m2/g are possible by dry decomposition of Ni oxalate while Ni citrate and Ni hydroxyacetate can achieve surface areas >200 m2/g. Theoretical yields, when starting from 500 mg Ni hydroxide Ni(OH)2 as precursor, are 403 mg for NiO or 316 mg for (metallic) Ni. Mass balances are seen to be almost closed in most cases when assuming NiO as final product with roughly 0.1–2% residual carbon left over from incomplete combustion of the organic counteranion. 3.2. Results for nickel—acid screening While dry powder decomposition only allows the synthesis of monometallic NiO, it is most desirable for catalytic applications to prepare well mixed multimetal oxides. Therefore, we resorted to wet combustion synthesis starting from an aqueous solution of dissolved metal salts in a combustible organic acid. For the modification of the traditional Pechini (citric acid combustion) method, our strategy was based on the selection of low molecular weight and easy to decompose multifunctional carboxylic acids such as diacids, hydroxoacids and ketoacids (as a replacement for the traditional citric acid) to (a) solubilize the metal precursors by forming metal carboxalte solutions and (b) to allow for low calcination temperatures to burn off the organics without inducing sintering of the fresh metal oxide surfaces formed.

135

Table 2 High surface area NiO–organic acid screening Acid

Calcinations (8C)

BET (m2/g)

Glyoxylic Glyoxylic

290 300

337 378

Glyoxylic Diglycolic Tartaric Oxamic Oxalic Citric Malic Lactic Malonic

325 325 325 325 325 325 325 325 325

269 245 205 189 185 153 149 108 99

Ketoglutaric Malic Tartaric Glutaric Succinic

350 350 350 350 350

200 196 184 168 146

Ketoglutaric

375

119

Ketoglutaric

400

105

Table 2 summarizes screening results for various acids whereby only calcinations temperature (but not calcinations time which was fixed to 4 h in air) and acid amount have been varied without a global optimization of the full parameter space (which will be addressed later). Table 2 identifies glyoxylic acid as the most suitable dispersant in the low temperature calcinations regime 290–310 8C with surface areas >350 m2/g whereas ketoglutaric acid requires higher calcination temperatures in the range 350– 400 8C to burn off coke. 3.3. Results for nickel—glyoxylic acid Table 3 gives an overview of NiO samples prepared by the novel proposed glyoxylic acid method starting from Ni hydroxide as precursor. A clear green solution is formed upon gentle heating of Ni hydroxide in aqueous glyoxylic acid. The aqueous Ni glyoxylate solution thus formed is stable for at least 1 year. Ni hydroxide was the best precursor and superior to Ni acetate for highest surface areas to be achieved, while results with Ni nitrate were poor. We preferred Ni hydroxide over the carbonate or hydroxicarbonate because of the excessive foaming and gas evolution (coming along with spillover) of the latter ones when coming into contact with the

Table 3 High surface area NiO by glyoxylic acid method 500 mg Ni(OH)2 + dispersant

Aging

State

Calcinations

BET (m2/g)

Yield (mg)

10 ml 6.25% glyoxylic

None

Foggy soln Recalcined

300 8C/3 h 300 8C/0.5 h

337 331

460 403

10 ml 12.5% glyoxylic

None

Grn soln Recalcined

300 8C/3 h 300 8C/0.5 h

279 298

440 390

1 g Ni(OH)2 + dispersant

Aging

State

Calcinations

BET (m2/g)

Yield (mg)

10 ml 12.5% glyoxylic 13 ml 10% glyoxylic 14 ml 10% glyoxylic 10 ml 15% glyoxylic 12 ml 12.5% glyoxylic 10 ml 15% glyoxylic 12 ml 12.5% glyoxylic 7 ml 20% glyoxylic 7 ml 20% glyoxylic

None None None None None None None None None

Grn Grn Grn Grn Grn Grn Grn Grn Grn

320 8C/4 h 320 8C/4 h 320 8C/4 h 320 8C/4 h 320 8C/4 h 320 8C/4 h 320 8C/4 h 320 8C/4 h 315 8C/4 h

299 310 312 338 324 325 326 327 327

835 829 837 847 851 862 845 836 865

soln soln soln soln soln soln soln soln soln

Standard temperature ramp: 45 8C/4 h ramp/120 8C/4 h hold/2 h ramp/300 8C/3 h hold (theoretical yield is 403 mg NiO from 500 mg Ni(OH)2).

136

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

Table 4 High surface area Ni mixed oxides by wet combustion synthesis Precursor

Dispersant

Aging for

State

Calcination

BET (m2/g)

Yield (mg)

0.5 g Ni(OH)2 + 100 mg Mn(OAc)2 (93 Ni:7 Mn)

7 ml 3 M ketoglutaric

3 weeks

Grn gel/foam

350 8C/4 h

149

427

1 g Ni(OH)2 + 100 mg Mn(OAc)2 (96 Ni:4 Mn)

15 ml 3 M ketoglutaric

3 weeks

Grn gel/foam Recalcined Recalcined Recalcined

350 8C/4 h 400 8C/4 h 450 8C/4 h 500 8C/4 h

170 136 111 76

863 858 843 836

500 mg Ni(OH)2 + 200 mg Fe(OAc)2 1 g Ni(OH)2 + 65 mg Mn(OAc)2 + 100 mg (NH4)2Pd(C2O4)22H2O

6 ml 10% glyoxylic 7 ml 20% glyoxylic

None None

Grn solution Grn solution

300 8C/4 h 315 8C/4 h

404 326

1 g Ni(OH)2 + 6.5 ml 20% glyoxylic + 100 mg (NH4)2Pd(C2O4)22H2O

80 mg Ce(OAc)3

None

Grn solution

320 8C/4 h

321

160 mg Cr(OAc)3 150 mg Al(OAc)3

None None

Grn solution Grn solution

320 8C/4 h 320 8C/4 h

334 352

acid. When aging the Ni glyoxylate solutions (i.e. evaporating water solvent) for a couple of weeks in open vials at room temperature in the hood, volume shrinkage by a factor of about 10 occurred and a transparent glassy green gel was formed, however, we found no difference as far as surface area of the calcined NiO is concerned when calcining the aged gel or directly drying/calcining the solution. We used inexpensive and common precursors selected from only a handful of compounds (nitrates, hydroxides, carbonates, acetates, oxalates, acac). Best results for the Pechini method (which are reported in the tables) were achieved with hydroxides and carbonates, i.e. when starting from the metal carboxylates. Dry decomposition of solid Ni glyoxylate or Ni oxalate also works well while the (dry or wet) decomposition of Ni nitrate only gives poor results. 3.4. Results for mixed nickel systems The proposed modified Pechini method is most powerful for the synthesis of mixed oxides and superior to other methods such as precipitation (difficult to achieve intimate mixing for many metals simultaneously) or dry decomposition (suitable only for monometallic oxides). The metals are entrapped and immobilized in the gel or polymeric resin that the acid turns into upon evaporating the solvent (water) when drying and/or calcining the solution. Table 4 gives examples of the binaries Ni–Mn and Ni–Fe oxides. 3.5. Characterization of high surface area Ni oxide XRD identifies NiO as the major phase, possibly contaminated with very minor contributions of metallic Ni particles or Ni2O3

(Fig. 2, glyoxylic acid prep). Chemical analysis showed a residual carbon content of approximately 0.5% that helps stabilizing the oxidic matrix against sintering. 3.6. Catalytic application of high surface area Ni oxides Fig. 3 compares the hydrogenation activity of Ni–Ce–Pd with Raney-Ni (supplied from Fluka) for the slurry phase hydrogenation of octene to octane. Ce has been incorporated into the NiO matrix as stabilizer against sintering and traces of Pd allow to lower to reduction temperature of NiO to about 250 8C (gas phase H2 reduction of NiO to Ni prior to feeding the octene reactant). It is seen that the doped Ni oxides (after careful reduction) are competitive to commercial Raney-Ni with respect to hydrogenation activity. Although we report the catalytic activity only for Ni-Ce-Pd as a representative example, many more Ni samples (with W, Co, Mn, Fe, Cr, Al, Ce, Nb, Mo as binders and various Pd precursors for the promoter) were screened with similar results, i.e. many samples were competitive to Raney-Ni in octene hydrogenation. Ni loadings for all doped Ni samples are similar (above 90%) because the binder and promoters add up to only a few percent of the material. 3.7. Results for Rare Earth oxides The synthesis of high surface area Rare Earth oxides has been attempted by all three methods (dry and wet combustion as well as precipitation). In the following we will be presenting representative results for holmium oxide. For commercially available Ho oxide nanopowder (as supplied by Aldrich 637327) we measured a BET of 14.5 m2/g. The rationale behind the design of the experiments was to screen the commonly available precursors and identify the best dispersant and calcination conditions for each precursor. No mathematical algorithm or evolutional techniques have been applied. 3.8. Results for holmium—dry decomposition

Fig. 2. XRD of high surface area NiO prepared by combustion synthesis with glyoxylic acid as dispersant.

Table 5 summarizes results obtained by calcinations of dry powder precursors. Ho oxide substantially free of carbon is pink whereas brown or black material is indicative of larger amounts of residual coke. It is seen that BET >100 m2/g is easily possible by calcination of Ho acetate whereas poor results are obtained for Ho carbonate as precursor. Holmium acetate gives higher surface areas than the carbonate because Ho belongs to the block of RE metals which form infusible acetates (Dy, Ho, Er, Tm) where much higher surface areas are obtained than for the other RE (except cerium where high surface areas are also possible). The decomposition of the Dy, Ho, Er, Tm

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

137

Fig. 3. Conversion vs. reaction time for the slurry phase catalytic hydrogenation of octene to octane over (pre-reduced) high surface area Ni-Ce-Pd catalysts (Ce content in the range from 1 to 6 wt%, Pd content from 0.02 to 0.2 wt%) compared to Raney-Ni. Reaction conditions: T = 55 8C, m(cat) = 300 mg, V(C8H16) = 1.5 ml, V(CH3OH) = 50 ml.

acetates is believed to directly yield the oxides without passing through the carbonate intermediate which is very stable (recall that the RE are basic) and therefore difficult to decompose and require higher calcinations temperatures resulting in lower surface areas. For RE metals other than the block of infusible acetates, only moderate surface areas are obtained (by dry decomposition) regardless of the precursor (carbonate, acetate, oxalate, acac) because the precursor is decomposed into the carbonate which releases CO2 only at elevated temperature due to the basic nature of the metal (recall the extreme thermal stability of alkaline earth of La carbonate). 3.9. Results for holmium—combustion synthesis Table 6 summarizes some results for broad acid screening as well as parameter optimization for the best acids identified

(malonic, glutaric, succinic). It is seen that neither the traditional Pechini acids (citric, tartaric, malic) nor the novel ones (glyoxylic, ketoglutaric or diglycolic that gave superior results for Ni, Co, Fe, Mn, In, Sn) work well for Ho (and the other Rare Earths) resulting in brown material (i.e. require higher calcination temperatures for complete burn off of coke accompanied by surface area losses) whereas the simple diacids malonic, glutaric, succinic were found to work best resulting in pink (i.e. carbon-free) material already at relatively low calcination temperatures below 400 8C and decent surface areas in the range 80–120 m2/g. Fine tuning of acid/metal ratio and calcinations conditions are required for best results (highest surface areas). For the combustion synthesis of Ho oxides, the carbonate was used as precursor (rather than the acetate which resulted in highest surface areas upon dry decomposition, see above) due to the easy and complete reaction with the organic acid whereby CO2

Table 5 Holmium oxide by dry decomposition Calcination

BET (m2/g)

Yield

pwd pwd pwd pwd pwd pwd

325 8C/4 h 335 8C/4 h 350 8C/4 h 375 8C/4 h 400 8C/4 h 450 8C/4 h

133.5 131.1 133.0 119.0 113.8 98.8

520 mg 616 mg 640 mg 562 mg 545 mg 581 mg

None None

Pink pwd Pink pwd

500 8C/4 h 360 8C/2 h

70.1 140.6

400 mg pink 533 mg pink

969 mg Ho acetate

None

Pink pwd Recalcined

350 8C/2 h 350 8C/1 h

137.6 136.8

550 mg pink 528 mg pink

988 mg Ho acetate

None

Pink pwd

370 8C/1 h

125.0

565 mg pink

866 mg Ho acetate 904 mg Ho acetate

None None

Pink pwd Pink pwd

275 8C/4 h 300 8C/4 h

133.2 134.0

487 mg pink 499 mg pink

836 mg Ho carbonate 1 g Ho carbonate 1 g Ho carbonate 1 g Ho carbonate 1 g Ho carbonate 1 g Ho carbonate 1 g Ho carbonate 1 g Ho carbonate

None None None None None None None None

Pink Pink Pink Pink Pink Pink Pink Pink

350 8C/4 h 350 8C/4 h 375 8C/4 h 375 8C/4 h 400 8C/4 h 425 8C/4 h 450 8C/4 h 500 8C/4 h

26.9 31.0 26.5 41.5 37.2 40.7 44.3 39.7

Precursor

Dispersant

State

939 mg Ho acetate 1.1 g Ho acetate 1.17 g Ho acetate 1.026 g Ho acetate 1.01 g Ho acetate 1.1 g Ho acetate

None None None None None None

Pink Pink Pink Pink Pink Pink

807 mg Ho acetate 970 mg Ho acetate

pwd pwd pwd pwd pwd pwd pwd pwd

Temperature ramp: 55 8C/1 h ramp/120 8C/1 h hold/1 h ramp/400 8C/4 h hold in 40 ml vial.

433 mg 517 mg 513 mg 463 mg 494 mg 474 mg 461 mg 460 mg

pink pink pink pink pink pink

pink pink pink pink pink pink pink pink

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

138 Table 6 Ho oxide by wet combustion synthesis Precursor

Dispersant

1 g Ho carbonate

3 ml H2O + 2 ml 12.5% glyoxylic

BET (m2/g)

State

Calcination

Pink soln Recalcined Recalcined Recalcined Recalcined Recalcined

375 8C/4 h 400 8C/4 h 425 8C/4 h 425 8C/4 h 425 8C/4 h 425 8C/4 h

25.2 20.8 22.8 32.1 35.2 35.0

478 mg 461 mg 459 mg 442 mg 429 mg 429 mg

Yield

brown brown brown pink pink pink

1 g Ho carbonate

5 ml 1 M tartaric

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

27.9 13.8

491 mg brown 474 mg brown

1 g Ho carbonate

7.5 ml 1 M malic

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

20.4 13.5

491 mg brown 481 mg brown

1 g Ho carbonate

10 ml H2O + 600 ml pyruvic Pink soln Recalcined

375 8C/4 h 400 8C/4 h

19.3 8.9

484 mg brown 478 mg brown

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

18.9 9.9

473 mg brown 468 mg brown

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

34.1 35.7

468 mg pink 454 mg pink

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

17.6 18.5

472 mg pink 465 mg pink

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

20.0 3.9

592 mg brown 513 mg brown

1 g Ho carbonate

1 g Ho carbonate

1 g Ho carbonate

1 g Ho carbonate

10 ml H2O + 500 mg diglycolic

10 ml H2O + 500 mg ketoglutaric

10 ml H2O + 1 g ketoglutaric

10 ml H2O + 500 ml 70% Glycolic

1 g Ho carbonate

10 ml 1 M lactic

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

7.7 6.5

437 mg brown 478 mg brown

1 g Ho carbonate

10 ml 0.5 M oxamic

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

20.2 13.4

493 mg brown 488 mg brown

1 g Ho carbonate

5 ml 1 M glutaric

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

77.7 65.8

466 mg pink 437 mg pink

1 g Ho carbonate

5 ml 1 M glutaric

Pink slurry Recalcined

350 8C/4 h 360 8C/4 h

81.7 79.0

471 mg pink 458 mg pink

1 g Ho carbonate

10 ml H2O + 500 mg trans-aconitic Pink slurry Recalcined Recalcined

375 8C/4 h 400 8C/4 h 425 8C/4 h

15.7 11.9 11.9

502 mg brown 492 mg brown 479 mg brown

1 g Ho carbonate

5 ml 1 M citric

Pink slurry Recalcined Recalcined

375 8C/4 h 400 8C/4 h 425 8C/4 h

39.0 21.8 18.1

485 mg brown 464 mg brown 449 mg brown

1 g Ho carbonate

5 ml 1 M malonic

Pink slurry Recalcined Recalcined

375 8C/4 h 400 8C/4 h 425 8C/4 h

41.9 24.8 20.9

440 mg pink 457 mg pink 452 mg pink

1 g Ho carbonate

10 ml H2O + 500 mg glycine Pink slurry Recalcined Recalcined

375 8C/4 h 400 8C/4 h 425 8C/4 h

18.8 11.3 9.7

515 mg brown 492 mg brown 484 mg yellowish

Pink soln Recalcined Recalcined

375 8C/4 h 400 8C/4 h 425 8C/4 h

6.4 6.4 6.2

498 mg brown 484 mg brown 476 mg brown

Pink soln Recalcined Recalcined

375 8C/4 h 400 8C/4 h 425 8C/4 h

33.2 29.7 23.1

497 mg brown 488 mg brown 469 mg pink

1 g Ho carbonate

1 g Ho carbonate

10 ml H2O + 500 ml pyruvic

5 ml H2O + 5 ml 25% gluconic

1 g Ho carbonate

4 ml 2 M tartaric

Pink slurry Recalcined

375 8C/4 h 390 8C/4 h

40.3 35.4

489 mg brown 482 mg brown

1 g Ho carbonate

6 ml 1 M glutaric

Pink slurry Recalcined

375 8C/4 h 375 8C/2 h

85.5 81.6

470 mg pink 454 mg pink

1 g Ho carbonate

7 ml 1 M glutaric

Pink slurry Recalcined

375 8C/4 h 375 8C/2 h

87.0 83.4

473 mg pink 464 mg pink

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

139

Table 6 (Continued ) BET (m2/g)

Yield

Precursor

Dispersant

State

Calcination

1 g Ho carbonate

8 ml 1 M glutaric

Pink slurry Recalcined

375 8C/4 h 375 8C/2 h

86.2 83.1

473 mg pink 462 mg pink

1g 1g 1g 1g

4 ml 2 ml 8 ml 7 ml

Pink Pink Pink Pink

slurry slurry slurry slurry

375 8C/4 h 375 8C/4 h 375 8C/4 h 375 8C/4 h

87.4 86.8 82.8 81.3

478 mg 476 mg 407 mg 404 mg

Ho Ho Ho Ho

carbonate carbonate carbonate carbonate

2M 4M 1M 1M

glutaric glutaric malonic malonic

pink pink pink foam pink foam

1 g Ho carbonate

4 ml 1 M malonic

Pink slurry Recalcined

360 8C/4 h 375 8C/4 h

17.0 15.3

499 mg pink-brown 489 mg pink-brown

1 g Ho carbonate

5 ml 1 M malonic

Pink slurry Recalcined

360 8C/4 h 375 8C/4 h

18.0 15.8

493 mg pink-brown 488 mg pink-brown

1 g Ho carbonate

10 ml H2O + 500 mg succinic Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

15.2 11.3

472 mg brown 470 mg brown

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

67.5 59.4

464 mg pink 443 mg pink

Pink slurry

375 8C/4 h

70.4

479 mg pink

Pink soln Recalcined

375 8C/4 h 400 8C/4 h

17.3 17.0

380 mg pink 377 mg pink

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

8.4 7.1

504 mg brown 486 mg brown

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

10.5 7.5

498 mg brown 476 mg brown

Pink slurry Recalcined

375 8C/4 h 400 8C/4 h

9.1 8.6

500 mg brown 498 mg brown

1 g Ho carbonate

1 g Ho carbonate

1 g Ho carbonate

1 g Ho carbonate

1 g Ho carbonate

1 g Ho carbonate

10 ml H2O + 750 mg succinic

10 ml H2O + 1g succinic 10 ml H2O + 2g ketoglutaric

5 ml H2O + 500 ml methoxyacetic

5 ml H2O + 1 ml methoxyacetic

10 ml H2O + 2 g diglycolic

1 g Ho carbonate 1 g Ho carbonate

7 ml 10% oxalic 9 ml 10% oxalic

Pink slurry Pink slurry

375 8C/4 h 375 8C/4 h

28.4 27.3

481 mg brown 483 mg brown

1 g Ho carbonate

7 ml 10% oxalic

Pink slurry Recalcined

400 8C/4 h 425 8C/4 h

44.9 29.8

486 mg brown 479 mg pink

1g 1g 1g 1g 1g 1g 1g 1g 1g 1g

8 ml 8 ml 7 ml 7 ml 7 ml 7 ml 7 ml 7 ml 7 ml 6 ml

Pink Pink Pink Pink Pink Pink Pink Pink Pink Pink

slurry slurry slurry slurry slurry slurry slurry slurry slurry slurry

360 8C/4 h 350 8C/4 h 340 8C/4 h 330 8C/4 h 320 8C/4 h 290 8C/5 h 280 8C/6 h 270 8C/8 h 300 8C/3 h 300 8C/6 h

83.6 87.3 93.9 96.5 98.9 110.9 112.9 113.5 110.0 51.4

Pink slurry Recalcined Recalcined

350 8C/4 h 375 8C/4 h 400 8C/2 h

11.0 11.6 11.3

Ho Ho Ho Ho Ho Ho Ho Ho Ho Ho

carbonate carbonate carbonate carbonate carbonate carbonate carbonate carbonate carbonate carbonate

1 g Ho carbonate

1M 1M 1M 1M 1M 1M 1M 1M 1M 1M

malonic malonic malonic malonic malonic malonic malonic malonic malonic malonic

3 ml 3 M citric

403 mg 408 mg 411 mg 420 mg 426 mg 440 mg 428 mg 430 mg 415 mg 452 mg

pink pink pink pink pink pink pink pink pink pink

foam foam foam foam foam foam foam foam foam foam

510 mg brown 494 mg brown 487 mg brown

Temperature ramp: 55 8C/4 h ramp/120 8C/4 h hold/1 h ramp/375 8C/4 h hold in 40 ml vial.

is released and the metal carboxylate is formed in situ. Generally we found higher surface areas for the metal carbonate or hydroxide precursors because no additional counter radical (acetate, oxalate) would have to be removed upon calcinations. The nitrates generally showed poor performance. In other words, the choice of precursor depends on the synthetic method and is certainly one of the parameters to be optimized. 3.10. Results for holmium—precipitation Table 7 summarizes precipitation results for 3 different Ho precursors (acetate, chloride, nitrate) that were pumped at 78 8C into excess NMe4OH base. We prefer pH titration over precipitation at constant pH because previous scouting experiments

showed high surface areas were easy to achieve by pH titration without elaborate optimization and precise adjustment of pH. It is seen from Table 7 that Ho oxides of about 100 m2/g are possible by calcinations of the freshly precipitated hydroxide. 3.11. Summary rare earth oxides—general trends The highest surface areas achieved for the different lanthanides and synthesis methods (precipitation, wet and dry combustion) are summarized in Table 8. With the exception of Ce (where precipitation was found to be far superior to combustion synthesis and CeO2 >300 m2/g could be produced), it is seen that precipitation can afford roughly about 100 m2/g for any metal whereas (dry or wet) combustion gives inferior results for all Rare

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

140 Table 7 Ho oxide by precipitation pH titration

T (8C)

Calcination

BET (m2/g)

Yield (mg)

5 g Ho(OAc)3 + 70 ml H2O pumped into 50 ml 1 M NMe4OH + 50 ml H2O

78

400 8C/4 h 350 8C/4 h

98.4 100.7

1317 1310

5.7 g HoCl3 + 40 ml H2O pumped into 50 ml 1N NMe4OH + 50 ml H2O

78

450 8C/4 h 550 8C/4 h

84.8 76.6

1405 1405

5.78 g HoCl3 + 40 ml H2O pumped into 50 ml 1 M NMe4OH + 50 ml H2O

78

350 8C/4 h 400 8C/4 h

100.7 98.9

1465 1448

13.5 ml 1.1269 M Ho(NO3)3 + 16.5 ml H2O into 50 ml 1 M NMe4OH + 50 ml H2O

78

350 8C/4 h 400 8C/4 h

90.9 84.4

1419 1432

Precipitation/pH titration in 250 ml three-neck flask; masterflex pH controller; peristaltic pump; isolation of ppt by centrifugation in two 40 ml vials; 4 wash steps in 40 ml vials. Temperature ramp: 55 8C/4 h ramp/120 8C/4 h hold/1 h ramp/350 8C/4 h hold in 40 ml vial. Table 8 Highest BET surface areas achieved by precipitation, dry or wet combustion synthesis for the Rare Earth oxides Highest BET surface areas (m2/g) achieved for Rare Earth oxides Dry decomposition Ce 168 Pr 23

Nd 33

Sm 32

Eu 27

Gd 33

Tb 52

Dy 122

Ho 141

Er 148

Tm 152

Yb 21

Lu 36

Precipitation Ce 306

Nd 114

Sm 80

Eu 72

Gd 64

Tb 88

Dy 84

Ho 101

Er 123

Tm 94

Yb 100

Lu 98

Eu 26

Gd 65

Tb 54

Dy 114

Ho 114

Er 132

Tm

Yb

Lu

Pr 120

Combustion Synthesis (novel Pechini acids) Ce Pr Nd 58 Sm 103

solution during drying) because the wet combustion method makes use of a non-stoichiometric acid/metal ratio (usually higher than 1, i.e. molar excess acid). Interestingly, superior results were obtained by dry decomposition of Dy, Ho, Er, Tm acetates only

Earth besides the block of Dy, Ho, Er, Tm where surface areas higher than 100 m2/g can be achieved by (dry or wet) combustion. It is recalled that dry and wet combustion are different (although a dry metal carboxylate may be formed by crystallization from the

Table 9 Synthesis of multimetal oxides by soft combustion synthesis BET surface area (m2/g)

Composition of product (mol%)

Acid

Acid/metal ratio (mg/mmol)

(mmol/mmol)

Temperature (8C)

Time (h)

Co47 Co47 Co47 Co48 Co48 Co48 Co14 Co14 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13 Co13

Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Glyoxylic acid Ketoglutaric acid Ketoglutaric acid Ketoglutaric acid Oxalacetic acid Oxalacetic acid Oxalacetic acid Diglycolic acid Diglycolic acid Diglycolic acid Oxalacetic acid Oxalacetic acid Oxalacetic acid

417.4 208.7 104.4 420.2 210.1 105.1 150.5 151.0 147.6 149.7 147.8 149.1 148.5 145.8 118.2 236.5 354.7 118.2 236.5 354.7 118.2 236.5 354.7 70.9 94.6 141.9

4.5 2.3 1.1 4.6 2.3 1.1 1.6 1.6 1.6 1.6 1.6 1.6 1.6 1.6 0.8 1.6 2.4 0.9 1.8 2.7 0.9 1.8 2.6 0.5 0.7 1.1

300 300 300 300 300 300 300 300 300 300 400 400 400 400 400 400 400 400 400 400 400 400 400 400 400 400

4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4

134.4 140.7 141.3 132.9 125.1 73.3 150.4 137.7 159.3 161.0 74.6 69.8 66.0 48.3 148.8 145.9 148.8 191.1 83.7 38.8 71.0 45.5 37.9 199.8 192.2 180.9

Glyoxylic acid Glyoxylic acid Glyoxylic acid Ketoglutaric acid Ketoglutaric acid Ketoglutaric acid

46.0 138.1 230.1 73.1 219.2 365.3

0.5 1.5 2.5 0.5 1.5 2.5

300 300 300 300 300 300

2 2 2 2 2 2

132.4 156.3 135.1 130.2 134.8 127.3

Co7 Co7 Co7 Co7 Co7 Co7

In29 Ce24 In29 Ce24 In29 Ce24 Sb28 Ce24 Sb28 Ce24 Sb28 Ce24 In16 Ce70 Sb16 Ce70 Ge18 Ce69 Sn17 Ce70 Sn17 Y70 Sb16 Y71 In17 Y70 Ge18 Y69 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70 Sn17 Y70

Fe7 Fe7 Fe7 Fe7 Fe7 Fe7

Sn16 Sn16 Sn16 Sn16 Sn16 Sn16

Ce70 Ce70 Ce70 Ce70 Ce70 Ce70

Calcination

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

whereas the carbonates always resulted in low surface areas. On the other hand, for all other Rare Earth outside this favorable block of 4 metals, dry decomposition always led to poor results regardless of the metal precursor chosen (acetate, carbonate, acac). As far as wet combustion is concerned, best results were obtained with malonic or glutaric acid as dispersants, i.e. essentially carbon-free (yellowish, white or pink) material >100 m2/g for Sm, Dy, Ho, Er. Traditional Pechini acids or the novel ones that proved to be well suited for many main group or transition metals (glyoxylic, ketoglutaric, oxalacetic, diglycolic) were less good for the Rare Earths. 3.12. Excerpt of results for higher dimensional multimetal oxides (example: Co systems) We have previously synthezised a library of Ru–Co–Ce–Y quaternaries by the modified Pechini method using Ce(III) nitrate, Co(II) nitrate, Y(III) nitrate and Ru(III) nitrosyl nitrate as precursors in aqueous glyoxylic acid as dispersant and calcining the solution at 350 8C for 4 h in air [20]. Catalytic results for combined CO oxidation and VOC removal have been presented elsewhere [10,20,21]. We have then been investigating other mixed Co oxides for applications in other areas. Table 9 gives an excerpt of ternary and quaternary Co oxides prepared by soft combustion synthesis. 3.13. Summary—general trends The highest surface areas achieved for the different synthesis methods (precipitation, wet and dry combustion) are summarized in Table 10. BET surface areas >100 m2/g are possible for many metals using a variety of methods after calcination in the temperature range 300–500 8C [7,11]. Precipitation (of the hydroxides) was found to be far superior to combustion synthesis for La (145 m2/g versus 45 m2/g) and Ce (306 m2/g versus 168 m2/g) and most of the Rare Earth (except the Dy–Ho–Er–Tm block where wet as well as dry combustion resulted in the highest surface areas). Precipitation has not been applied by us to Ti and Zr where combustion synthesis gave poor results. On the other hand, the modified Pechini method is well suited and resulted in surface areas >200 m2/g for Ni, Co, Fe, Mn, Y, Sn and Nb. Best results have been obtained for glyoxylic acid, oxalacetic acid, and ketoglutaric acid. Glyoxylic acid was found to work very well for Ni, Sn, In, Mn, Fe [8,9,12–14] and ketoglutaric was best for Co while oxalacetic gave

141

high surface areas for yttria [15]. Interestingly, neither glyoxylic nor ketoglutaric produced high surface area Cu oxide but diglycolic was the best acid identified resulting in decent surface areas >80 m2/g. The proposed modified Pechini method did not work well for the traditional carrier materials Ti and Zr while for Nb glycolic proved to be superior to glyoxylic. For V and Mo, glyoxylic and oxalic gave similar results. These novel Pechini acids have a low decomposition temperature in common, in the temperature range of 300–350 8C, thus allowing relatively low calcination temperatures and concomitant mitigation of surface sintering. These acids mostly decompose into volatiles (e.g. ketoglutaric acid ! acetone + CO2) and are less prone to coking. For instance, the isolated powders in the case of Sn, Ce, In are yellowish and in the case of Ho, Er pinkish with the mass balance almost closed thus indicating the absence of larger amounts of coke deposits. On the other hand, the high boiling traditional Pechini acids citric and malic yield black or brown materials when calcined <350 8C and would require temperatures of the order of 400 8C to burn off coke completely. The solubility of, for instance, Sn(IV) acetate, In(III) acetate, Mn(II) acetate or Ni(II) hydroxide in aqueous glyoxylic acid is very high thus allowing high productivities (space-time-yields) and rendering the ‘glyoxylic acid method’ a convenient and flexible tool for the synthesis of these porous metal oxides. Moreover, as Table 9 reveals, mixed metal oxides with high surface areas can also be synthesized, even in the presence of (hard to stabilize) base metals (such as Co in Table 9). The optimal calcination temperature is governed by the ability of the metal to form carbonates and their stability. The more acidic metals such as Ni, Co can be calcined at 300 8C whereas Zn requires slightly higher temperature of 325 8C to decompose Zn carbonate that might be formed as intermediate. The more basic Y has to be calcined at 400 8C for carbonate decomposition. For the strongly basic metals forming extremely stable carbonates upon combustion in the presence of organics (La), one has to resort to precipitation (by hydroxide bases) when high surface areas are wanted. For ceria, precipitation resulted in far higher surface areas >300 m2/g than combustion synthesis [11,18]. For the remaining Rare Earth metals poor results are usually obtained by combustion due to the high stability of the carbonates (as precursors or intermediates) with the exception of Dy–Ho–Er–Tm when decomposing their non-fusable acetates or slurrying their carbonates into aqueous malonic or glutaric acid [16].

Table 10 Highest BET surface areas achieved by precipitation, dry or wet combustion synthesis for metal oxides

142

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

As far as the noble metals are concerned, high surface area oxides >180 m2/g could only be produced for Ru (e.g. from Ru nitrosyl acetate and ketoglutaric acid) while the other precious metals got reduced by the organic acid to the metal (i.e. metallic state) upon mild heating already. Interestingly, for Ce and Ru, even the chloride precursors gave decent surface areas, possibly by in situ Deacon oxidation of residual chloride during calcination in air. 4. Discussion A variety of synthesis techniques have been used to prepare metal oxide materials. The most common techniques include conventional precipitation, the Pechini, or citric acid combustion process, and a variety of sol–gel techniques. Typical precipitation methods utilize stable, acidic metal salts in solution. The solution is combined with a base that increases the pH of the metal salt solution and destabilizes the metal salts to form metal hydroxides and/or metal carbonates that precipitate out of the solution. This reaction results in counter-anions of the metal salt, such as nitrates or chlorides, and the counter-cations of the base, such as Na, K, or NH4 being present. After the precipitation, it is usually desirable to remove the ions from the base and the salt by washing, usually with a solvent such as water. However, this does not typically remove all of the impurities. In order to avoid the ion contamination issue, precipitation with urea or hydrazine (which both decompose into volatiles upon boiling the solution) have been found to give comparable results to the use of other bases, such as NaOH or Na2CO3 [12,23,24]. Hydrazine or urea can be advantageous, since the precipitation agent is almost completely removed leaving little or no countercations. Hydrazine decomposes upon boiling into nitrogen, hydrogen and water, and the anion of the metal precursor (such as a chloride) is also removed from the system as a volatile gas, such as HCl. Urea breaks down to ammonia and CO2 with the ammonia released being the actual base/precipitation agent thus forming NH4Cl or NH4NO3 salts that may partly evaporate and partly reside in the solution. The solutions have to be heated to about 90 8C or refluxed during precipitation and aging thus adding to the energy cost. The Pechini, or citrate method [1–6], involves combining a metal precursor with water, citric acid and a polyhydroxyalcohol, such as ethylene glycol. The components are combined into a solution which is then heated to remove the water. A viscous oil remains after heating. The oil is then heated to a temperature that polymerizes the citric acid and ethyleneglycol by polycondensation, resulting in a solid resin. The resin is a matrix of the metal atoms bonded through oxygen to the organic radicals in a crosslinked network. The resin is then calcined at a temperature typically above 500 8C to burn off the polymer matrix, leaving a porous metal oxide. The Pechini method is advantageous in that it utilizes components that are inexpensive and easy to handle. However, the method results in materials having BET surface areas substantially lower than those materials created using precipitation and sol–gel methods. Typical sol–gel methods utilize metal alkoxide precursors in organic solvents with an aqueous inorganic acid, such as nitric acid or hydrochloric acid. The inorganic acid acts as a catalyst allowing the water to hydrolyze the metal alkoxide bonds in a hydrolysis reaction by protonation, forming a metal hydroxide and an alcohol. Subsequent condensation reactions involving the metal hydroxide units reacting with other metal hydroxide units or remaining metal alkoxides result in the metal molecules bridging, and water and alcohol being created. As the number of bridged metal molecules increases, agglomeration occurs, forming irregular agglomerates and eventually growing into a three-dimensional

amorphous polymer network, or a gel. The remaining water and alcohol, which is a neutral non-ionic unreactive organic solvent, is evaporated from the system leaving little to no traces of the former metal counter-anion behind. The gel is then calcined, resulting in a porous, solid metal oxide. While the current sol–gel processes produce porous metal oxide materials having surface areas superior to those produced by precipitation and the Pechini method, the method suffers from the alkoxide precursors being expensive, flammable, viscous and difficult and dangerous to handle. As can be seen from summary Table 10, high surface areas >100 m2/g are possible for many metal oxides and mixtures by precipitation, dry decomposition as well as Pechini and related methods. The various synthesis methods we have developed complement each other with respect to achievable surface area/ porosity and residual contaminations, in particular: Synthesis method

Advantage

Disadvantage

Precipitation

High surface area

Dry decomposition Pechini method

Easy and inexpensive Flexible; mixed metal oxide synthesis

Possibly Na, K, Cl contamination; metal losses Monometallic oxides only Residual carbon contamination

When screening alternative Pechini acids (different from citric and malic) we discovered that significant surface area gains for many main group and transition metal oxides can be realized by resorting to organic acids such as glyoxylic, ketoglutaric or oxalacetic acid that are easily decomposable at low calcination temperatures in the range 300–350 8C without concomitant formation of refractory coke. For the Rare Earth oxides, malonic acid and glutaric acid were identified as the best dispersants for Dy–Ho–Er–Tm (i.e. the block of non-fusable lanthanide acetates) with calcinations temperatures in the range 350–400 8C being optimal. Therefore, we propose a modification of the Pechini method using these novel acids as dispersants for the synthesis of fairly general classes of porous metal oxide as well as mixed metal oxides [7–18]. Often we observe gel formation when evaporating the solvent (either upon prolonged standing at room temperature or when heating up for drying). Whether the gel is composed of metal oxide bridges and/or a metal carboxylate network requires further investigation. Ketoglutaric acid, for instance, constantly decomposes (into acetone and CO2) upon stirring thus resulting in high losses of acid from the system until pH has risen to a point where gelation occurs. However, the mass balance (for the dried gel or in case of a too low calcination temperature) indicates considerable residual organics in the gel that must be burned off during calcination. In summary, the proposed new method is as easy and flexible to carry out as the classical Pechini method but allows the very high surface areas to be achieved that are characteristic of a sol–gel process and advantageously utilizes common inorganic metal precursors (rather than alkoxides). 5. Conclusion Over the past couple of years, we have synthesized several tens of thousands of samples on a gram scale by dry and wet combustion synthesis for various catalytic applications including high surface area carriers for WGS and emissions control (Ce, Y, Sn), solid acids (Nb), Fischer–Tropsch (Fe–Mn), hydrogenations (Ni), total and partial oxidations (Ru–Co, Mo–V), HDS and HDN (Ni–Co– Mo). Many high surface area metal oxides are available by a novel modified Pechini method (all water based soft combustion

C. Bluthardt et al. / Catalysis Today 137 (2008) 132–143

synthesis) using aqueous solutions of thermally labile organic acids as dispersants for common inorganic metal precursors such as formate, acetate, acac, hydroxide, carbonate, and even chloride that are cheap, commercially available, non flammable and nonalkoxide precursors. The concept consists of selecting easy-todecompose organic acids that allow low temperature calcinations thus preventing surface area loss by sintering. Glyoxylic, oxalacetic, and ketoglutaric acid decompose in the temperature range 300–350 8C into volatiles without excessive coking. Calcination temperature, acid/metal ratio, and net initial weight of metal precursor are the principal parameters for high surface areas to be achieved and must be carefully optimized in small increments. We have demonstrated that high surface areas are achievable by means of aqueous routes after careful optimization of synthesis parameters, without resorting to organic solvents, sol–gel, supercritical drying or templates/hydrothermal synthesis. Our combustion synthesis is also well suited for the preparation of mixed oxides from mixed metal solutions in aqueous organic acids. Productivity is high since concentrated metal solutions can be used and neither aging nor washing are required. High purity materials can be easily produced that are strictly free of Cl, Na, K by proper metal precursor selection and only contaminated by residual carbon. The proposed modified Pechini or ‘‘glyoxylic acid combustion method’’ is generally applicable to the acidic metals while for the strongly basic metals (alkaline earth, La, Ce) precipitation of the hydroxides is to be preferred due to the initial transformation of metal carboxylates into the very stable carbonates intermediates the decomposition of which would require very high temperatures thus resulting in pronounced particle agglomeration by sintering. The Pechini method is ideally suitable for high throughput robotic synthesis of metal oxides and mixed metal oxides in particular.

[3] [4] [5] [6] [7] [8] [9] [10]

[11] [12] [13] [14]

[15] [16] [17] [18] [19] [20]

[21]

[22]

References [1] Z. Hui, P. Michele, J. Mater. Chem. 12 (2002) 3787–3791. [2] B.L. Cushing, V.L. Kolesnichenko, C.J. O’Connor, Recent advances in the liquid phase syntheses of inorganic nanoparticles, Chem. Rev. 104 (2004) 3893–3946.

[23] [24]

143

M.P. Pechini, US 3,330,697, 1967. H.M. Reichenbach, P.J. McGinn, J. Mater. Res. 16 (4) (2001) 967. H.M. Reichenbach, P.J. McGinn, Appl. Catal. A 244 (2003) 101. US 6,372,686, US 6,352,955, US 5,977,017, US 5,939,354, US 2002/0042341, WO 97/37760 to Catalytic Solutions Inc. M. Schlichter, High Surface Area Metal Oxides, Diploma Thesis, Symyx Technologies, Santa Clara, 2005. A. Hagemeyer, et al., High surface area tin oxide catalyst carriers, in: 19th NAM Meeting, Philadelphia, PA, May 22–27, 2005. B. Smaka, et al., High surface area tin oxide catalyst carriers, in: 10th International Symposium on Catalyst Deactivation, Berlin, Germany, February 5–8, 2006. A. Hagemeyer, Z. Hogan, G. Streukens, H. Turner, A. Volpe, H. Weinberg, K. Yaccato, Novel Ru-Co catalysts for combined low temperature CO oxidation/ VOC combustion, in: ACS Spring National Meeting, Atlanta, GA, March 26–30, 2006. WO 2006/119311 (assigned to Symyx Technologies, 2006). A. Hagemeyer, Z. Hogan, M. Schlichter, B. Smaka, G. Streukens, H. Turner, A. Volpe, H. Weinberg, K. Yaccato, Appl. Catal. A 317 (2007) 139–148. A. Hagemeyer, M. Schlichter, A. Volpe, High surface area manganese and iron oxides, in: ACS National Meeting, Chicago, IL, March 25–29, 2007. A. Hagemeyer, M. Schlichter, A. Volpe, High surface area indium oxide and indium–tin mixed oxide, in: AIChE Spring National Meeting, Houston, TX, April 22–26, 2007. A. Hagemeyer, M. Schlichter, A. Volpe, High surface area yttria, in: 20th NAM Meeting, Houston, TX, June 17–22, 2007. C. Fink, A. Hagemeyer, M. Schlichter, A. Volpe, High surface area rare earth oxides, in: EUROPACAT VIII Conference, Turku, Finland, August 26–31, 2007. A. Hagemeyer, Z. Hogan, C. Lugmair, M. Schlichter, A. Volpe, High surface area niobia, in: ACS National Meeting, Boston, MA, August 19–23, 2007. C. Fink, A. Hagemeyer, Z. Hogan, A. Volpe, J. Yoder, High surface area ceria, Appl. Catal. A, in preparation. C. Brooks, S. Cypes, R.K. Grasselli, A. Hagemeyer, Z. Hogan, A. Lesik, G. Streukens, A. Volpe, H. Turner, H. Weinberg, K. Yaccato, Top. Catal. 38 (1–3) (2006) 195–209. S. Cypes, A. Hagemeyer, Z. Hogan, A. Lesik, G. Streukens, A. Volpe, H. Weinberg, K. Yaccato, Triple function Ru catalysts for low-temperature CO oxidation, in: VOC Combustion and SCR-DeNOx, Paper P44313, Proceedings AIChE Annual Meeting, San Francisco, CA, November 12–17, 2006. S. Cypes, A. Hagemeyer, Z. Hogan, A. Lesik, G. Streukens, A.F. Volpe Jr., W.H. Weinberg, K. Yaccato, High throughput screening of low temperature CO oxidation catalysts using IR thermography, in: J. Margitfalvi (Ed.), Combinatorial Chemistry & High Throughput Screening CCHTS, Special Issue on Catalysis, in press. Y. Chang Kang, S. Bin Park, Y. Whan Kang, Nanostruct. Mater. 5 (7/8) (1995) 777– 791. N. Sergent, P. Gelin, L. Perier-Camby, H. Praliaud, G. Thomas, Sens. Actuators B 84 (2002) 176–188. L. Urfels, P. Gelin, M. Primet, E. Tena, Preprints CAPoC6, Brussels, October 22–24, 2003, O20, pp. 203–212.